Soccerment Joins MLS Innovation Lab with AI Analytics

OK, let's unpack this. Today's deep dive is taking us

onto the pitch, specifically into the world of football or

soccer as many of you know it, and exploring how cutting edge

tech and AI are really changing the game.

We're focusing on one company in particular, Soccer Mint, and

looking at how their whole approach to data and analytics

is shaking things up, including, you know, their recent selection

for a Major League innovation program.

That's right, Soccer Mint. They really position themselves

as a sports tech company squarely focused on using AI

data analytics mainly for performance analysis, player

tracking in soccer, right? Their mission, as they stated,

is basically to speed up the adoption of predictive

analytics, you know, across all levels of the sport.

And that's kind of the mission for this deep dive, too.

For you listening, We want to give you a shortcut sort of to

understanding this whole revolution.

We'll explore their specific tools, look at how they're

measuring performance in totally new ways, ways that go, you

know, way beyond just goals and assists and also what their

collaboration with the league, like MLS means for the future of

football. So yeah, get ready to see the

game through a pretty fascinating new lens.

So let's dive into the tech itself.

What are Sockerman score solutions?

What allows them to actually do all this?

Well, their offerings really center around two key things.

First, getting the right data and 2nd making sense of it.

OK, for the data collection part, their main product is

called Exceed. Exceed and these are like smart

shing guards, right? Exactly.

These aren't your old school shing guards.

Not at all. They're embedded with tech

designed to collect really high quality, granular performance

data directly from the player. During games training both.

During matches and training sessions, yeah.

OK, so players wear these and they're capturing more than

just, you know, impact protection.

What kind of info are we talking about?

It's sort of a dual layer of data they capture.

You got the athletic metrics, things like toll distance

covered, Yeah, top speed, number of sprints, acceleration.

Standard stuff. Yeah, standard, but important.

But critically, they also record and geolocate technical events

right there on the pitch. What does that mean?

Geolocate technical events. It means knowing precisely where

and when a pass was made, where a cross was delivered from,

where a shot was taken, and even the power behind that shot.

Wow. Yeah.

That ability to tie the technical actions, the what to

the location, the where, with that kind of precision, that's a

really significant advancement in data collection right at the

source. OK, capturing the what and the

where of technical actions alongside the physical effort.

That sounds complex. What kind of tech is actually

packed into something worn on a player's shin?

Must be pretty small. It's quite sophisticated, yeah,

especially for such a small and, you know, durable form factor.

They contain multiple sensors, accelerometers, gyroscopes for

tracking movement, a magnetometer and a high

frequency GPS antenna operating at 20 Hertz, which helps capture

that location data precisely. That makes sense.

All of that plus a Bluetooth module to transfer the data,

It's all integrated into a device that weighs less than 100

grams per shin guard under 100 grams.

Impressive. And they still have to meet, you

know, strict safety standards, personal protective equipment

rules, CE certification for strength, which meant a pretty

significant development process. I've got.

Yeah, including partnerships for things like the frame and the

light guide. They mentioned working with

Protolabs for injection molding, which apparently allowed them to

do rapid prototyping and refine the design quickly.

Right. That attention to the physical

product, the durability, it really highlights the challenge

of getting reliable data in, well, a pretty harsh sports

environment. Absolutely.

So who is this X seed system really for?

Is it just for the absolute elite top tier clubs with huge?

Budgets. Well, while it's definitely

valuable for elite teams, X Seed was actually developed

specifically to try and make this kind of cutting edge tech

more accessible. Accessible to who?

To coaches, teams, clubs, even below that very top level, the

data gets delivered through a pretty user friendly mobile app

and a web dashboard. Offers clear visualizations,

player rankings, even includes some virtual coaching features

that use the data to give tailored feedback and advice.

Huh. Making advanced data more

attainable for a wider audience? That feels, well, pretty

significant for developing talent across the board, doesn't

it? I think so, yeah.

You mentioned 2 core solutions though.

What's the second piece of their puzzle?

That would be their X value platform.

This is the advanced analytics engine.

It's where the data collected by Exceed or maybe integrated from

other sources gets processed. Right, the brains of the

operation. Kind of, yeah.

It's where the AI and machine learning really come into play,

turning all those raw data points into hopefully actionable

insights for clubs, academies, players, ultimately supporting

more advanced predictive analytics.

OK, so exceed captures the Super detailed data X value makes that

data intelligent. This leads us perfectly into the

next question. Why do we even need these new

ways to measure performance? We've had goals and assists

forever. Why the big push for these

advanced metrics? That's really the core problem.

Soccer men and others in this space are trying to solve

football. You know it's inherently a low

scoring game. It's true.

And the outcome of key events like whether a specific shot

goes in or not can involve a pretty high degree of

randomness. Luck basically.

Right the deflection, the incredible save.

Exactly. So evaluating players or team

performance based only on those rare, sometimes random events,

it can lead to bias conclusions. Well, a player might just have a

lucky streak right outperform their underlying ability for a

short period. Or, conversely, a player could

be consistently creating absolutely brilliant chances

that their teammates just aren't finishing.

OK, so traditional stats don't capture that nuance.

Not fully. Clubs need more objective, more

robust measures. Things that aren't swayed so

easily by short term results or that kind of volatility.

And this is where probability and machine learning step in to

smooth things out. Precisely By using really

detailed event data, the context, the location, the type

of every action, these machine learning models can move beyond

just the binary goal or no goal outcome.

They could assign a probability to an action based on what

typically happens in similar situations across thousands,

even millions, of historical data points.

So it gives a measure of the quality of the action.

Exactly the underlying performance quality, rather than

just the final result which might have been lucky or

unlucky, it's much more objective.

OK, let's unpack some of these specific metrics Sacrament uses

then. Expected goals or XG.

That's probably the one people have heard of most.

What does it represent in their system?

Right XG expected goals. They define it as a pre shot

probability. It's all about quantifying the

quality of a scoring chance before the shot is even taken.

Before the shot, OK. Yeah, it assigns A probability

to each shot, estimating how likely it is to result in a

goal. And this probability comes from

a statistical model trained on hundreds of thousands of pass

shots. What factors does it consider?

Various contextual things. Shot location on the pitch.

That's usually the biggest factor.

Makes sense, closer is better. Generally, yeah.

Also the type of pass that led to the shot, the game state, the

body part use head, foot, the pattern play.

Things like that. So fundamentally, it's telling

you how good the opportunity was regardless of who took the shot.

Exactly right. Think of a shot from say just

inside the penalty area, pretty central that might get a value

around foot one XG meaning meaning that on average about

10% of shots taken from that specific situation typically end

up as goals. OK but it's really vital to

remember XG measures chance quality.

It's not predicting if that specific shot will go in, it's

not really measuring the individual finishers skill at

that point. It's real power as a predictive

tool comes when you aggregate lots of shots over time, like a

whole season or multiple seasons.

Then you can see if a player or a team is consistently

generating high quality scoring opportunities.

Got it. And how do they handle penalties

in XG? Penalties are usually handled

separately or assigned a fixed value.

Because there's such a unique high probability event,

Suckerman's model assigns a fixed .78 XG for penalties based

on the historical average conversion rate.

OK, .78 makes sense. So that's XG measuring chance

quality? What about expected assists or

XA? What's that metric capturing?

Expected Assists XA tries to measure the probability of a

pass becoming a goal assist. The whole idea is to credit the

players who are creating scoring chances for their teammates

through their passing ability. The creative players.

Exactly. But this metric, well, it's seen

a couple of different approaches in the analytics world.

Oh, different ways to calculate it, yeah.

A lot of models use what's called a shot centric approach.

Basically, whatever the XG value is of the shot that follows a

pass, that XG value gets credited back to the passer as

their XA. OK, seems logical.

It does, but Sockerman uses a different method, a pass centric

approach which they argue gives a more nuanced attribution of

credit. Pass centric, how does that

differ? OK, so Sockerman's model is

trained on all completed passes, not just the ones that

immediately lead to a shot. It evaluates whether the pass

itself successfully moved the ball into a dangerous area of

the pitch. Regardless of what the receiver

does next. Exactly, regardless of whether

the receiver shoots or dribbles or passes.

Again, they believe this gives a fair reflection of the passers

contribution to creating danger. Can you give an example to make

that clear? Yeah, they often use this goal.

Romulu Lukaku scored for Inter Milan.

Ivan Parisage played a fairly simple long pass from deep

inside his own half up towards Lukaku near the halfway line.

Lukaku. They went on this amazing solo

run, beat several defenders and scored a goal.

The initial XG value of Lukaku's shot itself was around .11.

Now in a typical shot centric model, Parisage would likely get

the full .11 XA for that initial pass.

Because it led to the shot. Right, but Sockerman's pass

centric model looks at the pass itself.

It recognizes that Parisage pass, while successful, didn't

actually put Lukaku directly into a high probability scoring

position. It was Lukaku who created most

of the danger after receiving. It So what XA did their model

give Parisage? Of very low value, just .001 XA.

Wow, big difference .11 versus .001.

Huge difference and they argue that .0001 value much better

reflects Parasix actual contribution to that specific

goal, where the value is really added by Lukaku's individual

skill after the pass. Their XA model, like the SG 1,

is a statistical model trained on millions of completed passes.

That distinction really is key, isn't it?

Measuring the inherent danger of the pass itself.

OK, so we have XG for shooting opportunity quality, XA for

creative passing quality. How do you combine those to look

at a player's overall offensive contribution?

Is there a metric? For that, yes, that's where

they're metric called expected offensive value added or OVA

comes. In OVA it basically combines a

player's non penalty XG and their XA, the chances they

create themselves through shooting and passing, and then

it subtracts the XA they received from their teammates.

So the formula is like non penalty XG plus XA generated, XA

received. Pretty much, yeah.

The idea is to try and isolate that player's individual impact

on creating scoring chances. How effectively they turn

possession into dangerous situations.

Exactly. It highlights how much a player

contributes to chance creation through their own actions, both

shooting and passing. But again, like XG and XA, this

metric is focused on chance creation, not whether those

chances actually get converted into goals.

Right, the process, not the immediate outcome.

And it also really relies on having that pass centric XA

model for it to work correctly and not double count value.

They pointed out, for example, that in the 20/20/21 season,

across the top 7 European leagues, Lewis Muriel at Atlanta

had a really high OVA, something like .73 per 90 minutes.

Indicating a strong individual offensive impact, creating lots

of chances himself. OK, so those metrics, XGXAXOVA,

they're largely focused on what happens before the ball is

struck or the pass outcome. What about evaluating what

happens after, particularly for shooting skill and goalkeeping?

Right, that takes us to expected goals on target or X got TT.

X got T, another expected metric.

Yeah. So while XG is that pre shot

probability, assuming an average shot placement, if it hits the

target, X got T is a post shot probability.

It's calculated only for shots that actually are on target.

And what extra information does it use?

It incorporates the original XG value, the quality of the

initial chance, plus the precise location where the ball crosses

the goal line or where the keeper saves it within the

frame. So it factors in the shot

placement. How does that help us understand

finishing or goalkeeping then? Well, the difference between the

exigot value of a shot and its initial XG value gives you

something called shooting goals added or SGA.

Shooting goals added. Yeah, this metric is

specifically designed to measure finishing scale.

It quantifies how much a player's shot placement actually

increase the probability of scoring compared to what an

average player might achieve from that same initial XG

opportunity. So players who consistently play

shots in corners or difficult spots for the keeper would show

high SGA. Exactly.

Sockerman's analysis has highlighted players known for

precise finishing, like, you know, Lionel Messi or Ciro and

mobile. They tend to demonstrate high

SGA per XG, meaning they consistently score more goals

than expected, even after accounting for the quality of

the initial chances. Because they're finishing is

just that good. OK, that makes sense.

For evaluating strikers, how does X dot T apply to

goalkeepers? It's used directly to evaluate

keepers through the goals prevented metric.

This is calculated pretty simply.

You take the total X dot T that a goalkeeper has faced from all

the shots on target against them, and you subtract the

number of goals they actually conceded from those shots.

So ex Gotti face minus goals conceded equals goals prevented.

Exactly. It provides a much more

objective measure of pure shot stopping skill.

It shows whether a keeper is saving more shots than an

average keeper would be expected to save given the difficulty in

placement of the shots they faced.

Who stands out there? Yano Black is often the prime

example highlighted in these discussions.

Their analysis credited him with preventing something like almost

35 goals between 20/17/18 and the time of the analysis, and he

was preventing about .28 goals per ex guy T faced in that

period, which is really impressive.

Wow, that's a powerful way to quantify a goalkeeper's value

beyond just clean sheets focusing on the difficulty of

the saves. OK.

Any other crucial passing metrics we should cover?

Yes, one more important one is expected passes or expected.

Passes OK, similar concept to the others.

It's a model trained on millions of passes, and it calculates the

probability of any given pass being successfully completed.

Based on what? Based on factors like its

length, its direction, the positions of players around it,

game pressure, various contextual factors.

But hang on, we already have simple pass completion

percentage stats. Why do we need X pass?

What does it add? That's a great question.

It adds context about difficult. Simple completion percentage

treats all completed passes equally.

Right. A5 yard sideways pass counts the

same as a 40 yard diagonal ball through traffic.

X pass tells you how difficult the attempted passes actually

were. It helps identify players who

aren't just completing lots of easy passes, but who are

consistently completing the more difficult, riskier passes at a

higher rate than you'd expect on average.

How do you see that in the numbers?

You see it when a player's actual number of completed

passes is significantly higher than their total X pass value

for all the passes they attempted.

They're basically outperforming the expected completion rate

based on the difficulty profile of their passing.

Any examples? Players like Tony Cruz or Marco

Varadi are often cited as classic examples of midfielders

who excel here. They consistently show an

ability to successfully execute challenging progressive passes

more often than the average player would.

Cruz for instance I think of the 20/20/21 season should something

like a 5.19% over performance compared to his ex pass value.

That's fascinating. It really highlights the

difference between, say, a player who just keeps possession

with safe passes versus one who actively tries and succeeds at

making those difficult line breaking passes.

Exactly. Gives you a much better sense of

their true passing ability under pressure.

Okay, this is all incredibly insightful about on field

performance. Let's shift gears a bit.

How does all this detailed data and these advanced analytics

impact the business side of football, specifically thinking

about player recruitment? Yeah.

This is where it gets really interesting from a market

perspective, because despite the enormous amounts of money

sloshing around at football, the transfer market is still

characterized by, well, significant inefficiencies.

There's still considerable room for arbitrage.

Arbitrage meaning finding value that others miss.

Essentially, yeah, identifying and acquiring undervalued

players. And this doesn't just apply to

potential transfer fees, although it does there too.

It applies very clearly to player salaries as well.

So you're saying clubs could potentially find players who are

performing at a really high level based on these objective

metrics, but who are earning significantly less than other

players with comparable performance?

Absolutely. Their analysis suggests exactly

that. There are players on relatively

modest salaries who are performing at levels comparable

to or maybe even exceeding players earning exponentially

more money. Why does that inefficiency still

exist? You'd think in such a high

stakes, increasingly data aware environment, performance and

salary would align more closely. Well, Sockerman points to a

couple of key reasons. A primary one they suggest is

that data analytics, while definitely growing it, isn't yet

universally or perhaps deeply enough integrated across all

clubs decision making processes. So not everyone is using it

effectively yet. Right.

Or consistently enough to fully link objective performance data

to salary negotiations and valuations across the board.

They also mentioned identifying some structural issues within

the market itself. Even for players who are

supposedly in their prime years, say between 25 and 30, where you

might expect the market to be more efficient and settled.

Apparently it's not always the case.

Interesting. So for a club that is willing to

fully embrace this data-driven approach, it potentially

provides a real competitive advantage in recruitment.

Precisely. They frame it as arguably the

single clearest way for clubs to consistently gain an edge in the

market, right? Now how so?

Well, by having access to better, more objective

information through things like Exceed and the advanced metrics

from XW, and crucially, having the analytical tools and

expertise to extract actionable intelligence from it, clubs

could potentially make much smarter recruitment decisions.

They can capitalize on opportunities that opponents

relying more on, say, traditional scouting or

subjective judgments, might completely miss.

OK. And how does Soccer Mint

specifically use their data to try and identify these hidden

gems or undervalued players? What's their method?

They approach this by directly comparing a player's performance

against their net salary. For performance, they use their

own proprietary metric, the Sacrament Performance Rating, or

SPRS. PR.

Importantly, this SPR metric is adjusted for the specific league

level the player competes in, so that allows for, you know,

better like for like comparisons across different competitions.

Right, because performing well in one league isn't the same as

another. Exactly.

And then for the salary data, they integrate information from

external sources, specifically mentioning capology, which

tracks football finances. So visually, are they just

plotting performance on one axis and salary on the other looking

for outliers? Essentially, yes.

They plot players as PR against their salary, and they often use

a logarithmic scale for salary. Which makes sense because the

range of wages in football is just enormous, right?

From modest earners to superstars.

Right, a log scale compresses that huge range.

So what are they looking for on that plot?

They're looking for players who fall into that potential sweet

spot, high up on the performance scale, a high SPR, but

relatively low down on the salary scale.

The bargains, basically. The potential bargains, yeah,

but they have a more formal way to quantify undervalued beyond

just eyeballing the chart. How do they do that?

First, they segment players by position Goalkeeper, defender,

midfielder, forward. Because you need to account for

ositional differences in both erformance profiles and tyical

salary ranges. Makes sense?

Then, using statistical techniques like robust linear

regression, they establish a typical trend line for each

position. This line basically shows the

expected salary for a player given their SPR within that

position group. OK.

So the line represents the market rate for performance.

Kind of, yeah. The general relationship between

performance and pay for that position.

Undervalued players are then identified as those whose data

points fall significantly below that established trend line.

How? Far below.

They actually quantify that vertical distance below the line

and normalize it into an undervalue index, often indexed

to 100 for easy comparison. They focus their search

particularly on players. Their algorithm flags as true

outliers, often defined as having an SPR above a certain

threshold, say 50, and a high undervalue index score.

And did their deep dive into players aged 2530, the prime

years confirm this inefficiency exists?

Yes, they explicitly stated that their analysis of that age group

confirmed that plenty of this market inefficiency does indeed

exist. It's not just about finding

young, unproven talent. That's a really clear

methodology. Have they seen players

identified using this approach actually make significant moves?

Does it work in practice? Yes they have.

They've highlighted specific examples of players they flagged

as undervalued hidden gems based on this SPR versus salary

analysis who subsequently moved to larger clubs, like Who They

mentioned Steven Berquise moving to AX and Pierre Lee's Malou

transferring into Nord City as examples of players fitting that

profile who made moves after being identified.

That's pretty powerful validation for their approach,

finding inefficiency and seeing it potentially corrected by the

market later. Exactly.

And they've also tried to make this more accessible for their

own users now. They mentioned integrating the

salary data directly from capology into their analytics

platform for subscribers to their analyst pluses tier.

So clubs using their platform can run these kind of undervalue

analysis themselves. That seems to be the idea, yeah.

OK, that's fascinating. Which brings us right back

around to the news we touched on at the very start.

Sockerman was selected for the MLS Innovation Lab.

How does all this connect? Is this why MLS pick them?

It absolutely connects. I mean, being selected for the

MLS Innovation Lab is, it's a direct recognition of the

potential value that Sockerman's technology and analytics,

everything we've just discussed, bring to the sport.

It's a big deal getting into something like.

That Oh, definitely. They were chosen for the

program's second cohort, the one starting in 2025, after what

sounds like a pretty rigorous global evaluation process.

MLS said. They reviewed hundreds of

companies. Wow.

And what is the MLS Innovation Lab specifically trying to

achieve? What's its purpose?

The program is designed by MLS itself to help shape the future

of sports, basically, and also drive the league's own growth.

They do this by identifying, testing and then nurturing

promising startups and advanced technologies.

And MLS has specific areas they're focusing on.

Yes, they defined key priority areas that they want this cohort

of companies to help enhance. Those are fan engagement, media

technology and critically for soccer men on field player

performance. Right, which is exactly soccer

men's wheelhouse, as we've been discussing.

Precisely. It's a perfect fit on paper.

So getting selected gives soccer men a direct pathway to actually

test their tools exceed X value within the real MLS environment.

Exactly that. It provides soccer men with a

really significant opportunity to test and showcase their

innovations in real world MLS settings right across the

league's ecosystem. What kind of setting?

They mentioned things like major youth development events, MLS

Knickers Tea Fest, the Generation Agitus Cup, the MLS

Knickers Tea Cup. These are big events with lots

of teams, lots of players. It offers a great platform to

prove the value and scalability of their tech at a pretty large

stage. That does sound like a

monumental step for a company like Soccer.

Mint it. Certainly seems like it.

Aldo Commie, who's Soccer Mint CEO, described it as a

significant milestone for them. He also expressed pride in being

part of the cohort and interestingly noted they were

the first Italian company ever selected for the program.

First Italian company, huh? And what was the perspective

from the MLS side? What did they say?

Chris Schlosser, who's the SVP of emerging Ventures at MLS,

commented on it. He acknowledged the success

they'd already had with the first cohort of the lab and

expressed real excitement about collaborating with Soccer Mint

and the other companies in this new group.

His quote specifically mentioned the goal is to Co create new

ideas and push the boundaries in the game of soccer.

Co create new ideas that sounds very collaborative.

What are the potential outcomes for companies like soccer meant

that go through this innovation lab, what happens at the end?

Well, the program culminates in opportunities to really present

their capabilities. Companies in the cohort might

get to present directly to MLS executives, even team owners, at

an event called the Future of the Game Showcase.

When does that happen? It's typically held during the

MLS All Star Game activities, so high visibility.

Definitely. And what could come from that

exposure? Potentially it could lead to

long term strategic partnerships either with the league itself or

with individual MLS clubs, and maybe even potential investment

opportunities down the line. It effectively provides a

pathway for potentially much deeper integration into the

whole MLS structure. It really underscores MLS's

broader strategic focus on tech and innovation, doesn't it?

Fits right alongside things like their big partnership with Apple

for MLS season pass or their docu series work.

Absolutely. It shows they're actively

looking to integrate advanced technology, pretty much all

facets of their operation from what happens on the field to the

front office decision making and how they engage with fans.

OK, so bringing this whole deep dive full circle back to you

listening right now, think about the impact of this technological

revolution happening in football.

How does this big shift towards detailed data, advanced

analytics, how does it change how you might follow your

favorite team or how you evaluate individual players, or

maybe just how you think about the game itself?

Yeah, it's clearly moving beyond just counting goals or assists,

isn't it? Towards these much more nuanced

probabilistic measures of performance and potential.

Definitely. We've really explored how

companies like Sockerman are using that incredibly detailed

data collection through tools like the XE Shing Guards and

then applying advanced machine learning models in platforms

like XVU to create all these metrics.

We talked about XG XE XE, Xova XE, Goody TX Pass, which aim to

provide much more objective insights into player performance

and contribution. And we saw how that data can

then be leveraged potentially to identify those market

inefficiencies in player recruitment.

Yeah, finding the hidden. Gems.

Well, ultimately touching the eye of major leagues like MLS

and leading to big opportunities like their Innovation Lab

program. It's a fascinating conversions

of sport, tech and business. Absolutely.

And maybe that leaves us with one final thought for you to

consider. As technology continues to

integrate ever more deeply with sports, influencing everything

from training and tactics to scouting and player evaluation,

how might the very definition of talent OR value in a player

continue to evolve? How much will it be shaped by

this increasing depth and maybe increasing accessibility of

data? What stands out to you the most

about this whole data revolution that's transforming football?

Soccerment Joins MLS Innovation Lab with AI Analytics

About this episode